Understand this update in under a minute
- In one sentence OpenAI’s April 22, 2026 launch of workspace agents in ChatGPT lets organizations build shared agents that run across tools and Slack under admin controls, with research preview availability and credit pricing starting May 6, 2026.
- What happened OpenAI dated the workspace agents announcement April 22, 2026 and describes them as an evolution of GPTs powered by Codex.
- Why it matters Workspace agents can run in the cloud with access to files, tools, and memory, and they can be scheduled or deployed in Slack.
- What to do next OpenAI states workspace agents will be free until May 6, 2026, with credit-based pricing starting then.
Why it matters
The most valuable agent use cases are repeatable workflows: the same weekly report, the same onboarding checklist, the same customer follow-up. Shared, permissioned agents make that work scalable, but they also raise the bar for governance and monitoring.
What changed
The product shift here is from one-off chat to durable automation. OpenAI positions workspace agents as shared assets that can gather context, take actions across connected tools, and continue working when you’re away—especially inside Slack where teams already coordinate.
Practical read
For operators, the key details are controls and observability: who can build agents, what tools they can access, which actions require approval, and how analytics and version history are handled. Those are the features that determine whether agents reduce risk or increase it.
What to watch
The May 6, 2026 pricing transition is also a forcing function: teams should measure real ROI now, pick a short list of high-leverage workflows, and decide what reliability and review standards are required before wider rollout.